Outline

Ingegneria Sismica

Ingegneria Sismica

A Study on an Early Warning System for Identifying Illegal Dumping of Construction Waste and Debris Near Power Construction Sites Based on Remote Sensing Images

Author(s): Zilong Zhou1, Tingming Ye1, Wenliang Lin2, Peng Li1, Yulong Lyu1, Jianmin Bai2
1Jibei Electric Power Research Institute, State Grid Jibei Electric Power Co., Ltd., Beijing 100045, China
2State Grid Jibei Electric Power Co., Ltd. Zhangjiakou Power Supply Company, Zhangjiakou, Hebei, China
Zhou, Zilong. et al “A Study on an Early Warning System for Identifying Illegal Dumping of Construction Waste and Debris Near Power Construction Sites Based on Remote Sensing Images.” Ingegneria Sismica Volume 43 Issue 2: 1-18, doi:10.65102/is2026753.

Abstract

 For solving the difficult problems that relate to the unlawful pouring of building rubbish and fragments close to electric power construction places—including dispersed objects, broken dividing lines, big mixing with open building places and lawful treating areas, and slow reacting speeds in traditional inspections—this article puts forward one recognition and early warning system which is based on remote sensing. This system has built a multi-source data arrangement frame which includes an engineering restriction layer, a time remote sensing layer, and a high-resolution detailed check layer. At first, it carries out the definition of the candidate search region on the basis of power line corridors, tower foundations, substation expansion areas, construction access roads, and already approved disposal places. It then uses constraint-guided spatiotemporal boundary refinement networks to perform anomaly screening, three-branch feature coding, and target boundary restoration.Finally, a hierarchical early warning mechanism is established by integrating criteria such as persistence, expansion, deviation from legality, and environmental sensitivity. Results show that CSBR-Net achieves a peak F1 score of 89.2% on the overall test set, with an average F1 of 89.6% across scenarios, and maintains an 87.4% recognition rate even under wet surface conditions; the latency of the complete model is 43.0 ms per image, representing improvements of 4.6 and 7.3 percentage points over the optical baseline in F1 and mIoU, respectively. Typical examples have shown that risk numerical values can pass the alarm and attention critical points before the goal region obviously enlarges, hence giving early notification for on-the-spot checking. This research therefore points out that putting engineering law limit requirements, time change data, and fine divided boundary description into one single decision-making chain can effectively promote the stability and deployable ability of illegal soil and waste dumping inspection and early warning systems that lie near electric power construction projects.

Keywords
remote sensing imagery; power construction; dumped soil and debris; illegal dumping identification; early warning system

Related Articles

Qianwen Xiong1, Yuhong Chen1
1Guangzhou University of Chinese Medicine, School of Pharmaceutical Medicine, Guangzhou,Guangdong,China,510006
Zhihao Jiang1,2, Limi Chen1,2, Jing Yang1
1Hainan Vocational University of Science and Technology, Haikou 571126, China
2Institute for Mathematical Research, Universiti Putra Malaysia, Serdang 43400, Malaysia
Limi Chen1,2, Zhihao Jiang1,2, Jing Yang1
1Hainan Vocational University of Science and Technology, Haikou 571126, China
2Institute for Mathematical Research, Universiti Putra Malaysia, Serdang 43400, Malaysia
Hui Yuan1, Minjie Chai2, Siqing Xu1, Jinsong Li1, Jinwan Zheng1
1Electric Power Research Institute, State Grid Shanxi Electric Power Co., Ltd., Taiyuan, 030001, Shanxi, China
2Jincheng Power Supply Branch, State Grid Shanxi Electric Power Co., Ltd., Jincheng, 048000, Shanxi, China
Yanhan Zhu1,2
1China Academy of Cultural Heritage, Chaoyang District, 100029, Beijing, China
2Beijing University of Civil Engineering and Architecture, Xicheng District, 100044, Beijing, China